Professional Documents
Culture Documents
The Analysis of Biogenic Amines
The Analysis of Biogenic Amines
Thesis
Presented for the Degree o f
Master o f Science by Research
By Brian O Sullivan B.Sc.
Under the supervision o f Dr. Rosaleen Devery and Dr. Michael O Connell.
February, 2000
I hereby declare that this material, which I now submit for assessment on the
programme o f study leading to the award o f Masters degree by Research
described is entirely my own work and has not been taken from the work o f
others save and to the extent that such work has been cited and acknowledged
within the text o f my work.
Brian O Sullivan
ID No.:
Acknowledgements
I would also like to thank Dr. Dermot Diamond, Dr. Thersa Grady and Richie
Maguire for their help.
TABLE OF CONTENTS
Abrviations
List o f Tables
ii
List o f Figures
iii
Abstract
Chapter 1.
1.1
Introduction
1.2
1.2.1
Non-Fermented Foods
1.2.1.1
Fish
1.2.2
Fermented Foods
11
1.2.2.1
Cheese
11
1.2.2.2
W ine
13
1 .3
Producti on by Micro-organisms
15
1.4
Amine Oxidases
16
1.5
'
18
1.5.1
Fluorometric M ethod
18
1.5.2
22
1.5.3
1.5.4
ELISA
25
1.5.5
26
1.5.6
Biosensors
27
Chapter 2
32
2.1
2.2
Fluorometric Assay
2.3
33
34
Production in bacteria
34
2.4
Microbiological media
35
2.5
Bacterial strains
36
2 .6
histamine-producing bacteria
36
2.7
37
2.7.1
37
2.7.2
38
Chapter 3
40
3.1
Introduction
41
3.2.1
3.2.2
41
3.3.1
42
45
3.3.2
46
3.3.3
48
3.3.4
51
3.4
Identification o f a histamine-producing
Microorganism
51
3.5
54
3.6
Discussion
54
Chapter 4
59
4.1
Introduction
60
4.2
Results
61
4.2.1
65
4.2.2
67
4.3
Discussion
67
Chapter 5
70
5.1
Introduction
71
5.2
Results
74
5.3
Discussion
76
References
Appendix
Chromatograms
80
Abbreviations
AO AC
Cis
Carbon 18
DAO
Diamine oxidase
DCU
ELISA
GC
Gas Chromatography
HPLC
LB
Luria Bertani
MAO
monoamine oxidase
mM
Millimolar
MRS
NaCl
Sodium chloride
Na 2C 0 3 Sodium carbonate
NaOH
Sodium Hydroxide
NH 4
Ammonia
nm
nanometer
OPT
o-phthaldialdehyde
ppm
PVC
TFIB
TLC
TMA
trimethyl amine
UV
ultraviolet
List o f Tables
Table 1.1
Table 1.2
Table 1.3
Table 1.4
Table 1.5
Table 1.6
Table 1.7
Table 3.1
Table 3.2
Table 3.3
Table 3.4
Table 3.5
Table 3.6
Table 3.7
Table 3.8
Table 3.9
Table 4.1
Table 5.1
Table 5.2
Table 5.3
Table 5.4
Colour
generation
from
positive
histamine-producing
List o f Figures
Figure 1.1
Figure 1.2
Figure 1.3
Figure 1.4
Figure 1.5
Figure 1.6
Figure 1.7
Figure 1.8
The presence o f an antigen is detected using the enzymelabelled antibody. The concentration o f coloured product is
proportional to the concentration.
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8
Figure 3.9
Figure 4.1
Figure 4.2
Figure 4.3
Standard
Curve
Determination.
for
Dansylated
Histamine
by
HPLC
Figure 4.4
Figure 4.5
Figure 4.6
Figure 5.1
The
structure
of
the
chromogenic
ionophore
Representation
of
the
deprotonation
of
lithium
Figure 5.4
Figure 5.5
iv
Abstract
Biologically active amines in cheese and fish arising from metabolic activities o f
food-borne microorganisms have been implicated as the causative agents in many
food poisoning outbreaks. An awareness o f amine levels in foods today is therefore
important in relation to food spoilage and safety.
In recent years there has been increased consumer awareness about food
composition and safety and a corresponding increase in regulatory action. The food
industry requires reliable and cost effective analytical methods for process and
quality control to meet the needs o f the consumer. Therefore, this study focussed on
comparing conventional analytical methods involving HPLC and fluorimetry, for
histamine analysis in food.
Irish cheeses and canned tuna samples were selected for analysis o f histamine
content by a standard fluorometric technique. The cheeses, Cheddar, Cooleney,
Edam, Emmental and Brie were analysed over a three-week period and observations
of any changes in histamine levels were made. All samples were stored at 4C until
analysis to replicate the retail outlet storage conditions. The histamine levels found
in cheese in this study were low in comparison to levels in reported literature,
ranging from 0.2 to 4.3 mg/lOOg and were non hazardous for consumption. Canned
tuna was analysed by the same technique. Results from this study showed that the
tuna contained high levels o f histamine, 20 mg/lOOg that increased to hazardous
levels upon putrefaction after a 96 hour period. A HPLC method o f detection was
established based on the derivitisation o f histamine with dansyl chloride and UV
detection with the aim o f detecting histamine quantitatively in food samples and to
facilitate correlation studies with the fluorometric method o f detection. A novel
method, based on an amine oxidase system coupled to an ammonia sensing
calixarene, was investigated as an alternative and improved method o f histamine
detection.
1.1 Introduction
Biogenic amines are chemically defined as
decarboxylation
through
endogenous
(naturally
occurring)
Aromatic
Tvramine
Phenyletyylamine
Aliphatic
Putrescine
Cadaverine
Spermine
Spermidine
Heterocyclic
Histamine
Tryptamine
Primary amine:
R ---
Secondary amine:
N ------- H
H ----
R-
Tertiary amine:
R ---
N --------- H"
I
R'
H istam ine, 4-(2-ami noethyl )imidazo Ie) is a primary heterocyclic amine derived
from the decarboxylation o f the amino acid L-Histidine (Figure 1.2). It plays an
important role in biological processes (vasodilation and gastric acid secretion)
but also occurs exogenously in the food supply and can cross the intestinal
barrier. Intoxication can result if significant quantities cross the barrier and enter
the bloodstream (Taylor, 1988). There have been many reported incidences o f
food poisoning involving histamine/other amines and as a consequence
determination of these levels in foods is an important aspect o f food safety.
I
yw
'
H2 KH-CH2 '
H2 H-(Cff2 > 2
IST1DIE
H3STARINE
\
i 2B
---------
SUfTWIiKE
UTSiME
EOjH
0 2 H
H2*KH-{CH2>3-NN2
ORNITHINE
ARGININE
I
H fcH -C C H sV ^
mRE5C]/1E
2 h
* 2 #-CFK'Ctf2 )r M 2
h2
H2 -(CH2 )^C=N
H2 N- CH2>5-HH2
AGHATME
CADAVERINE
SPERMIDINE
I
2 - ia i2 3 - f l- (Cl? v - f O b ) 3-HH2
SPEWUIE
1.2.1.1 Fish
Scombrotoxin poisoning (histamine poisoning) is caused by the ingestion o f
foods containing high levels o f histamine and may also include other vasoactive
compounds such as cadaverine and putrescine. Scombroid fish poisoning was so
called as it was historically associated with the ingestion o f spoiled fish from the
scromboid families which include tuna, mackerel, skipjack and bonito. Nonscromboid fish such as mahi-mahi, herring, sardines and anchovies have also
been implicated with outbreaks. (Morrow et al. 1991). These fish all contain
high levels o f free histidine and other amino acids in the fish muscle, which
serves as an excellent medium for the growth o f invading bacteria. (Karmas,
1981). Proteolysis, either autolytic or bacterial, may play a role in the release o f
free histidine from tissue proteins. (Taylor, 1986). Histidine can be catabolized
in two ways in fish muscle: Amino-acid deamination to obtain urocanic acid or
histidine decarboxylation to produce histamine. (Santos, 1996). Any food with
the appropriate amino acids and that is subjected to certain bacterial
contaminants and growth may lead to scombroid poisoning if ingested. (Food
and Drug Administration, 1998).
The illness is an intoxication, so the incubation period is rather short ranging
from immediate to 30 minutes after ingestion o f the spoiled food. (Taylor, 1986)
The duration o f the illness is usually three hours but may last several days (Food
and Drug Administration, 1998). Histamine poisoning is most often confused
diagnostically with food allergies as they share identical symptoms and can be
treated by antihistamines. It can be distinguished from food allergy on the basis
of; a) no previous history o f a reaction to the food, b) high attack rate in group
outbreaks and c) high histamine levels detected in incriminated foods. (Taylor,
1986, Taylor, 1988).
Symptoms include flushing, nausea, vomiting, diarrhoea, headaches, dizziness,
rash and sometimes swelling o f the face and tongue (Morrow, et al. 1991).
There has been considerable doubt as to whether histamine is the causative
agent in scombroid fish poisoning. Evidence supporting this doubt is the fact
that when individuals are administered histamine orally no significant
symptoms associated with scombrotoxism are observed. Luthy and Schlatter
(1983) in a placebo-controlled double blind experiment showed that histamine
(25mg) administered orally in apple juice to 27 healthy volunteers did not show
any significant effect. This was also true o f wines containing natural amounts o f
histamine (non-detectable to 21 ppm). (Luthy and Schlatter, 1983).
However Morrow et al. (1991) provided evidence that histamine is the causative
toxin o f scombroid fish poisoning. Poisoned individuals who had ingested fish
containing high levels of histamine began to exhibit symptoms ten to thirty
minutes after ingestion. Their urine was analysed one to four hours afterwards
and showed histamine and N-methylhistamine levels o f 9-20 times and 15-20
times the normal mean, respectively. Levels dropped with time and after 14
days they were back to normal. It was concluded that the histamine in the urine
was most likely derived from the spoiled fish and the results showed histamine
to be the toxin responsible. (Morrow et al. 1991).
Two hypotheses have been proposed to explain this paradox between the
toxicity o f histamine when consumed in conjunction with spoiled fish and the
lack of toxicity when reagent-grade histamine is ingested. The biogenic amines
cadaverine and putrescine, which have been shown to be present in spoiled fish
(Mietz and Karmas 1977), acting as potentiators, are at the basis o f both
hypotheses. Since the oral ingestion o f toxic levels o f histamine alone does not
lead to fatality, it was proposed that a barrier to histamine absorption exists in
the gut. Large amounts o f mucin bind to histamine hindering its absorption
A new problem has been the increased duration o f fishing voyages to obtain
greater catch quantities and less than adequate handling and storage facilities
onboard ships which results in less fresh fish for the processor and increases the
potential for histamine formation. (Karmas, 1981).
Canned fish are o f considerable concern in relation to food safety. The fish is
prepared from previously frozen fish, which is then thawed before processing
and thus is subjected to additional handling, which may result in higher
histamine levels. Karmas and Mietz showed that histamine exhibited a marked
increase in concentration with canning where levels were twice that o f uncanned
fish. (Karmas and Mietz, 1978).
Product
Temperature
0F (-17.8C)
32F (0C)
38F (3.3C)
40F (4.4C)
50F (10C)
70F (21.1C)
90F (32.2C)
Safe Shelf
life(days) with
Rapid Cooling
No limit
14
Safe Shelf
life(days) with
Delayed Cooling
No limit
10
7
5
7
3
As can be seen from Table 1.2 the safe shelf life is significantly reduced above
4.4C and fish should not be subjected to this temperature for more than four
hours.
Source
found
Acineiobacter Iwojfi
0, 15
Aeromonas hydrophelia
0, 15
B eef
Citrobacter freundii
15
Skipjack tuna
Clostridium perfringens
15,30
Skipjack tuna
Enterobacter sp.
30
Food, tuna
Escherichia coli
15,30
Tuna
Hafnia alvei
15, 30
Tuna, mackeral
Klebsiella sp.
15
Tuna, mackeral
Morganella morganii
0,15,30
Scombroid fish
Proteus vulgaris
30
Proteus sp.
15
Vibrio sp.
15
Mackeral
Incidence
Fish is the major source o f protein in the Japanese diet and therefore it is not
surprising that this country has a very high incidence o f food poisoning
outbreaks.
Most outbreaks in Japan involve a large number o f people with the largest
outbreak occurring in 1973 where 2656 people were effected. Given that raw
fish is the preference in Japan one would expect it to be responsible in the
majority o f cases, however most illnesses are associated with cooked fish. One
argument to explain this is that only the highest quality fish is used for
consumption in this manner.
Since the 1970s the countries with the most reported cases o f histamine
poisoning have been Japan, U.S.A. and Britain but better reporting may account
for such figures.
The type o f fish caught and methods o f harvesting are important determinants in
histamine poisoning. In Scandinavian countries where there is a high
consumption o f fish, one would expect many outbreaks like Japan but there
have been very few incidents o f poisoning. The type o f fish consumed is not
prone to histamine formation and catching and storage temperatures are low
which decrease the possibility o f such formation.
10
Fish
Histamine
Cadaverine
Putrescine
source
(mg/IOOg)
(mg/IOOg)
(mg/IOOg)
Canned
1.97
0 .8
0 .1 2
Kim and
Bjeldanes
tuna
Canned
Reference
118
1 0 .8
1.25
1979
0.38
0.15
0 .1 2
M ietz and
tuna *
Canned
Karaias
tuna
Canned
25.3
1.93
0.25
(1978)
tuna*
Canned
Taylor,
3.46
1978
tuna
Sardines
0.79
1.2.2.1 Cheese
After fish, cheese is the next food product responsible for food poisoning
outbreaks, and numerous cases have been reported. (Sumner et al. 1985).
Cheese acts as a perfect environment for amine production supplying substrates,
the presence and activity o f bacteria and enzymes, proteolysis, water availability
and ideal ripening and storage conditions. (Edwards and Sandine, 1981). The
amine content in fresh milk is quite small which is why proteolysis may play
11
12
Cheddar 1
Histamine
Tyramine
(mg/lOOg)
(mg/lOOg)
26.5
108.5
Reference
De Vust et al.
1976
Cheddar 2
4.6
65.0
Edam
8 .8
Trace
Camenbert
2.4
Trace
Brie
1.4
39.8
Cheddar
14.0
24.0
Voigt et al.
Edam
N.D.
31.0
1974
Camenbert
7.0
1 2 .0
Gouda
7.5
29.0
Cheddar 1
5.8
Chambers
Cheddar 2
1 .2
and
Camenbert
0.7
Staruszkiew icz
Swiss
116
1978
1.2.2.2 W ine
13
Wine
Level (p.g/L)
Red wine
Level (f-ig/L)
White wine
Cuvee 1989
3,776
Riesling 1989
120
Bordeaux 1989
2,197
Riesling 1988
42
Zweigelt 1991
281
Chardonnay
35
1988
Goldeck 1988
Messwein
133
1991
Cuvee 1987
92
Langenloiser
1988
14
Amino acid decarboxylases are not widely distributed among the bacterial
population but species such as Bacillus, Proteus, Klebsiella, Pseudomonas,
Salmonella
and the
causing food poisoning. O f the organisms tested only Proteus morganni and
Enterobacter aerogenes had the capability o f causing an outbreak o f food
poisoning.
trypticase-soy broth-histidine
medium, which is generated from raw tuna fish. There are a lot o f differences
between the media and solid tuna flesh but the growth and histamine formation
in the media should demonstrate an organisms potential to cause histamine
accumulations in fish. The average histamine production for P. morganni and E.
aerogenes following 7-hour incubation amounts to 500 and 133mg/100g tuna,
which is enough for illness to occur. (Taylor et a l 1978a).
As discussed in section 2.1.1 low storage temperatures are used in the fishing
industry to control bacterial histamine formation. The lower temperature limits
for production o f toxicologically significant levels o f histamine in tuna fish
infusion broth were 7C for Klebsiella pneumonia; 15C for both Proteus
morganni strains and 30C for Hafnia alvei, Citrobacter freundii and
Escherichia coli. The abilities o f the Proteus and Klebsiella species to produce
significant levels o f histamine at these low temperatures is critical to the
freezing and storage conditions on board fishing vessels.
P. morganni and K. pneumoniae produced large quantities o f histamine in a
short period o f time (<24 hours) with K. pneumoniae producing a maximum
amount (42^moles/ml by 24 hours o f incubation at 37C and then declining.
The levels o f histamine production with P. morganni accumulated up to 72
hours incubation. This is a significant result as fish containing these strains may
15
accumulate high levels o f histamine in a short amount o f time but yet show no
appearance o f spoilage. (Behling and Taylor, 1982).
Detection o f histidine-decarboxylating bacteria may be difficult as often they
account for a minority o f bacterial species present in fresh fish. To facilitate
detection,
was developed
for the
1 2 0 0 nmole
histamine per millilitre. Other amines will not interfere in the assay. (Sumner
and Taylor, 1988).
16
RCHO + H 20 2 + NH 3
The enzymes are divided into two separate groups. The first group is the flavin
co-factor dependent enzymes known as monoamine oxidases (MAO), EC
1.4.3.4. These oxidases are located in the outer mitrocondrial membrane and
metabolise neuroactive amines. Inhibitors o f these enzymes are effective in the
treatment o f Parkinsons disease, schizophrenia and clinical depression.
The second group is the copper amine oxidases or diamine oxidases, EC 1.4.3.6 ,
which are found in animal tissue and plasma, plants, yeasts, fungi and bacteria.
(Malmstrom et al. 1975). The metabolic function o f these enzymes is the
breakdown o f a number of biologically active amines.
Highly purified enzymes have been obtained from the fungus Aspergillus niger,
pea seedlings, bovine blood plasma, pig plasma and pig kidney cortex.
The amines most rapidly oxidised are the aliphatic monoamines with chain
lengths C 3-C 6 including agmatine and histamine. Tyramine and tryptamine are
oxidised at a slower rate. Pig kidney oxidases oxidise alkyl diamines such as
cadaverine and putrescine most rapidly but histamine is also a good substrate.
The purification o f the enzyme from this source has been used in a number o f
studies (Bouvrette et al. 1997) because o f its specificity for these diamines
which are very important from a toxicological point o f view.
The optimal substrate concentration for enzyme activity is 34mM for cadaverine
and Im M for histamine in the pH range 6.3 and 7.4. Histamine inhibits activity
at concentrations greater than Im M and hydrogen peroxide produced during the
reaction o f a diamine and the enzyme inhibits activity at neutral pH. (Mondovi
etal. 1971).
Copper is the only metal found in significant quantity in the amine oxidases and
is an essential component for activity. The copper ions are firmly bound but can
be partially removed by treating with diethyldithiocarbamate (Yamada and
Adachi, 1971). Activity is lost on removal o f the copper and reacti vation can be
achieved by addition o f suitable amounts o f Cu .
17
18
M imicking o f histamine.
Suppression o f histamine fluorescence
Generation o f increased histamine fluorescence during the excitation o f the OPThistamine complex with daylight or UVA light.
19
m o l/i1
4
5
4
3
5
5
S p e rm id in e
N o ra d re n a lin e
S p e rm in e
P u ire sc in e
T ry p ta m in e
5 -O H -T ry p tam in e
A d re n a lin e
3 -O H -T y ra m in e
M eth y l h is ta m in e
A d e n in e
X
X
X
X
X
x
I0 -S
lO "4
1 0 -1
1 0 -3
10-3
1 0 -3
io-;
i o -2
2 X ID -2
6 X 1 0 -2
20
The major difficulty with histamine analysis in foods is the need to selectively
extract the compound from the biological matrix and eliminate all possible
interfering compounds. This is why the sample treatment step prior to the actual
fluorometric assay is so important. Taylor el al made some efficient and
improved modifications to the method o f Shore (1971) and that o f Rice el al
(1975). Firstly the initial extraction o f amines from the food was investigated.
Methanol proved more efficient compared to 10% trichloroacetic acid (TCA),
(Lerke and Bell, 1976) and 0.4M perchloric acid (Rice, 1975). Recoveries were
103,
86%
21
efficient
0.5
tyr
0.4
SM
03
HIS
CAD
0.2
0.1
ETA
*
0.1
*
0.2
'
0.3
PUT
SD
0.4
TRYP
05
06
22
23
and spermidine. These observations formed the basis o f the equation for the
quality index. (Mietz and Karmas, 1977).
Index =
Class 1: 0-1
good
Class 2: 1-10
borderline
Class 3:
decomposed
>10
There are several other methods based on derivitization with dansyl chloride for
the simultaneous determination o f biogenic amines in fish, cheese and beer and
have reasonably quick elution times o f 24 minutes or less. (Wei et al. 1995,
Buiatti et al. 1994, Vall and Malle, 1997 and Vall and Gloria 1997).
A group o f reagents (polymeric benzotriazoles) are known to derivatize most
amino acids/peptides rapidly and efficiently and previously were only used for
amino acid protection and peptide synthesis. Gao et al. (1998) developed new
polymeric reagents based on the polymeric benzotriazole with an o-acetylsalicyl
or 9-fluoroenyl labelling group that provided increased UV absorptivity and
fluorescent properties for derivatized amines. Derivatization is performed off
line (pre-column) at 60C for only 10 minutes and the limit o f detection is 1 to 2
picomoles for polyamines using the fluorenyl label and fluorescence detection.
(Gao et al. 1988).
chloroformate
(FMOC),
separation
by
reverse
phase
24
Figure 1.7: The presence of an antigen is detected using the enzymelabelled antibody. The amount of coloured product is proportional to the
antigen concentration.
Co Id urles
Substrate
. L .
Antigen
Antibody (> )
Labelled
Enzyme
with Enzyme ( E )
antibody
surrounded by
bound to
coloured
antigen
product
25
concentration in the sample, ie: the greater the concentration o f the antigen in
the sample the lower the amount o f labelled antigen binding to the antibody.
The sensitivity o f the assay is 0.8|ig/L.
Serrar et al, developed a monoclonal antibody-based ELISA based on the same
competitive system. In order to obtain antibodies specific for histamine it is
necessary to provoke an immune response from an animal against the
compound. However, histamine is a small molecule and alone is not
immunogenic. It must be coupled to a large immunogenic carrier molecule
usually a protein to stimulate antibody production. Once the monoclonal
antibodies (mAbs) were obtained and purified they were retested against
histamine-protein conjugates, other amines and negative controls to investigate
their antigenic specificity. It was found that they recognised histamine only and
showed high affinity to histamine after chemical derivitisation. The lower limit
o f detection was found to be lOng/ml. (Serrar el al. 1995).
H 2O 2 is then detected by the formation o f crystal violet from the leuco base in
the presence of oxidase.
Specificity o f the test: DAO also catalyses the breakdown o f other amines
associated with decomposition. The amines spermine and spermidine are found
to decrease with spoilage and are usually present in non-interfering amounts but
cadaverine and putrescine increase with spoilage, (Mietz and Karmas, 1977), so
they may add to positive reactions with histamine in the screening test.
However the procedure would fail to detect any non-histamine type spoilage in
fish.
Samples are absorbed onto filter paper strips by inserting them into a cut in the
dorsal are o f the fish for 5 minutes. The strips are then placed into reaction tubes
26
for colourimetric detection. The strip results were correlated with a standard
method o f detection by excising the same fish sample, homogenising and
analysing by the AO AC fluorometric method. Poor correlation was found which
was due to a broad range o f saturation intensities after 5 minutes on the strips,
which led to variations in sample size. The uneven distribution o f histamine in
tuna was also a contributing factor.
To overcome these problems, samples are usually taken from the area o f highest
histamine content, the gills and a smaller amount o f absorbent is used so that the
samples can become saturated quickly. (Lerke et al. 1983).
The sensitivity o f the test can be adjusted to test samples with a particular
histamine level by altering the horseradish peroxidase concentration.
1.5.6 Biosensors
Currently the food industry is not very receptive to biosensor technology but in
the long run it is hoped that a niche in the market can be developed. Biosensors
have to compete with other analytical methods in terms o f cost, performance
and reliability. Chromatography is the method o f choice for multiple sample
analysis and recently HPLC and GC instrumentation has been improving
making them more cost effective. The fact that they can analyse several
compounds simultaneously increases their reliability and speed o f analysis. Also
simple dipstick tests are easier and more user friendly than portable biosensors
for field testing and these account for why there is only limited acceptance o f
biosensors in the food industry.
The food industry spends 1.5-2% o f its total value sales on quality control and
appraisal. Recent trends in increased regulatory action and consumer awareness
about food safety has created a need for reliable and cost effective analytical
methods. New legislation requires the extensively labelling o f all major and
minor constituents present in the food. (Luong et al. 1997)
The functioning o f a biosensor involves the combination o f biochemical and
electrical interactions. It consists o f a receptor, which is an immobilised,
biologically responsive material linked in close contact to a suitable transducing
element.
27
The food industry has a need for simple, rapid and inexpensive methods that are
ideally automated. Process control in food is more complicated than other
industries
due
to
the
complicated
biochemical
matrices
involved.
Examples
Carbohydrates
Amino Acids
L-arginine, D-alanine.
Alcohols, Phenols
Gases
Amines, Amides
Inorganic Ions
28
The critical factor for biosensor design is maximal retention o f bioactivity and
biostability o f the biological molecule. By increasing these factors, a more cost
effective biosensor can be designed to compete with existing analytical
methods.
The AOAC fluorometric assay, HPLC or TLC has traditionally determined the
analysis o f biogenic amines in the food industry. However due to extensive
sample pre-treatment and time consumption, the methods are not as efficient as
desired.
Biosensors owing to their many advantages have become more and more
prominent as analytical devices in this industry and much research is being put
into further applications o f them.
Trimethylamine (TMA) is an important component o f the smell o f spoiled fish
and so lends itself as a good detection element for such spoilage. A sensor has
been developed for TMA detection based on an ammonia-sensing electrode,
which consists o f a glass (pH) electrode and a AgCl reference electrode in an
internal filling solution o f ammonium chloride, neutral salts and a dye. The
filling solution is separated from the analyte by a gas-permeable, ion-permeable
membrane. Any dissolved ammonia in the sample diffuses across the membrane
leading to an increase in pH, which is detected. In order to detect TMA, the
internal filing solution has been replaced with 0.01M TMA hydrchloride and
0.04M Potassium chloride. Formaldehyde is added to the sample solution in
order to decrease the response o f the electrode to ammonia. The analyte
concentration range is O.l-lOmM. (Chang et al. 1976).
29
t
Am ine oxidase
o f C 2-C 6 , benzylamine,
phenylethylamine,
histamine
and
agmatine. Aliphatic diamines, C 4-C 6, are oxidised at a lower rate. (Yamada and
Adachi, 1971).
The method correlated well with the official AO AC method, (AO AC, 1990) and
histamine recoveries were 100%. (Ohashi etal. 1994).
Male el al. (1996) also used this amine oxidase system in the development o f a
amperometric biosensor for determining histamine, cadaverine and putrescine.
The system utilised an amperometric electrode for the detection o f the product
hydrogen peroxide and was linear up to 6 mM with a lower limit o f detection of
30
25|j,M for the three substrates. The enzyme was a diamine oxidase purified from
porcine kidney, which exhibits specificity towards diamines like cadaverine and
putrescine but also deaminates histamine. (Male el al. 1996).
The influence o f biosensors will rise in the food and beverage industry in the
near future. Advances in improving the stability o f the biological component,
simultaneous analysis o f multiple analytes and mass production o f biosensors
will in the long run create a more cost-effective analytical device.
The aims o f the thesis are to study the different levels o f histamine in both tuna
and five varieties o f Irish cheeses by conventional methods o f detection. These
methods will be High performance liquid chromatography and fluorimetric
detection. The production o f histamine by food microorganisms will also be
monitored over time by the above methods. A new method o f detection utilising
a group o f compounds called calixarenes will be investigated and its potential
use as an analytical tool evaluated.
31
Chapter 2
Materials and Methods
32
100
ml
stoppered volumetric flask. The flasks were heated in a water bath for 30 min at
60C and allowed to cool before adjusting the contents to 100 ml with
additional methanol. The samples were transferred to capped polypropylene
tubes and centrifuged at 2000 rpm in an IEC PR-6000 centrifuge for
2 ml
1 /2 0
min. A
A 5 ml broth culture o f Providencia retgerri 865 was grown overnight and was
used to inoculate one 100 ml culture o f LB broth and one 100 ml L-histidine
supplemented culture o f LB broth. A growth curve was constructed by taking
O.D. readings at 550 nm at regular intervals. Simultaneously samples were
taken for histamine quantification by a fluorometric method as follows.
A 1.0 ml aliquot o f broth culture was added to 9.0 ml o f methanol and heated at
60 C for 15min and allowed to cool. 1.0 ml o f the methanol extract was run
through an ion exchange column (Dowex 1X-8 in the hydroxide form, 80 x 5
mm). The column resin was prepared as described (AOAC, 1980). The column
was washed with 35 ml of distilled water and the eluant was collected in a 50 ml
volumetric flask and the volume raised to 50 ml with distilled water. (Behling
and Taylor, 1982).
Flistamine content was determined using a fluorometric assay (Shore, 197 lb),
Column eluant (2 ml) was placed into a test tube followed by 0.4ml o f 3M
NaOH and the tube shaken. A 100 pi aliquot o f 1% (w/v) o-phthaldialdehyde
(Roth, 1971) was added, mixed and the reaction was let to proceed for 4 min
exactly. The reaction was stopped by addition o f 200 pi 3M HC1. Fluorescence
was measured at an excitation wavelength o f 360 mn and an emission
wavelength o f 450 nm (slit widths: excitation 10.0, emission 5.0).
34
2 .4
M ic ro b io lo g ic a l M e d ia
All media constituents were obtained from Oxoid unless otherwise stated.
M edia were solidified with 1.2% agar No.3 where necessary. D istilled water
was used in all preparations. Sterility was ensured by autoclaving at 151b/in2 for
20
minutes.
lOg
Yeast Extract
5g
NaCl (Sigma)
1Og
H20
1 Litre
pH 7.5
Tryptone
0.5%
Yeast Extract
0.5%
L-Histidine.2HCl
2.7%
NaCl
0.5%
C aC 0 3
0 . 1%
Agar
2.0%
Bromocresol purple
0.006%
pH 5.3
35
36
Histamine
dihydrochloride,
methanol,
diethyl
ether,
sodium
Method
Histamine dihydrochloride (165.7 mg) was dissolved in 10 ml o f de-ionised
water. This gave a concentration o f lOmg/ml.
To the histamine solution (50 (4,1) an aliquot o f 2M NaOH (1ml) was added,
followed by the addition o f 10 (j.1 o f benzoyl chloride. The solution was then
vortexed for 30 seconds and then left standing for a further 20 minutes.
Saturated sodium chloride (2ml) was then added, followed by extraction with
4mls o f diethyl ether.
37
Chromatographic Conditions
A gradient elution system was employed. The gradient elution program was set
at
1.1
(55:45, v/v) for 2.5 minutes. The program proceeded linearly to methanol: water
88:22 (v/v), with a flow rate increasing from 1.1 ml min ' 1 to 1.3 ml min ' 1 over
3.5 minutes. This was followed by the same composition and flow rate for 2
minutes, then decreased over 7 minutes to methanol: water (55:45 v/v) at 1.1 ml
m in 1. Finally the system was re-equilibrated at methanol: water (55:45, v/v) for
10
Derivatization Reaction
A concentrated solution o f histamine dihydrochloride (16.57 mg/100 mL water
and diluted
1 /1 0 0
38
washing the sides o f the flask thoroughly. The solution was then extracted with
three 5,0 ml portions o f diethyl ether, the ether layers were then combined and
evaporated using a rotary evaporator set at 60C. The residue in the flask after
evaporation was dissolved in acetonitrile and made up to a volume o f
1 0 ml.
From this stock solution, other dilutions were made up with acetonitrile (80, 60,
40 and 20%). A 20 pi sample o f each solution was injected into HPLC (Waters)
by a fixed loop. A 200 jj.1 aliquot o f extracted sample was treated in the same
way as the standards.
Chromatographic Conditions
Separating Column: Inertsil 50D S-2 (250mm x 4.6mm, 5|j,m)
Mobile Phase: Acetonitrile: Methanol: W ater (3:10:3)
Flow Rate:
1.2 mL/min
A,max:
254nm
39
Chapter Three.
Fluorometric Method of Analysis
40
3.1 Introduction
A simplified fluorometric assay for food analysis, as described by Taylor et al.
(1978b), was established as the standard method for determining levels o f
histamine in various foods. The method involved extraction o f histamine from
the biological matrix into methanol followed by protein denaturation and
precipitation and then further extraction into n-butanol. The organic extraction
step was selective for histamine and was dependent on the presence o f a salt
(sodium carbonate), its concentration and the type o f organic solvent used. Prior
to analysis it was important to assess the specificity and accuracy o f the method
and also to investigate possible interferences in the assay. The amount o f
histamine that is recovered from the sample clean-up process is an essential
factor and so spiking samples with known amounts o f histamine should show
the efficiency o f the procedure. A linear correlation between fluorescence and
histamine concentration is necessary in order to determine histamine levels
present in samples and this is demonstrated by Fig. 3.1 which shows a high
degree o f linearity with a R 2 value o f 0.997.
41
Histamine
Histamine +
Histidine (a)
Histamine +
Histidine (b)
Histamine,
histidine,
putrescine,
cadaverine,
glutamine and
trvptophan. (a)
Histamine,
histidine,
putrescine,
cadaverine,
glutamine and
tryptophan, (b)
Concentration
Measured
(ppm )
3.80
3.94
Recovery
3,91
103%
4.14
109%
4.03
106%
100%
104%
Two food samples (canned tuna) were extracted and treated according to the
method described by Taylor et al (1978b) and subjected to the fluorometric
assay. The assay was performed with and without the acidification step by 3N
hydrochloric acid and the results were read fluorometrically. Table 3.2 shows
the differences in amine levels in tuna samples.
Table 3.2 Levels of amines in tuna with and without acidification of the
OPT-histamine condensation complex.
- Acid
218.0 ppm
199.8 ppm
+ Acid
214.7 ppm
197.4 ppm
42
Fluoresence
Cone, (ppm)
Histamine
0
0.005
0.025
0.050
0.10
0.25
0.40
0.50
1.00
100.0
0
0.906
4.693
8.318
19.213
51.563
91.268
113.853
-
Histidine
(pre. col.)
0
0.182
0.6115
1.77
2.771
7.156
12.076
12.236
-
Histidine
(post col.)
0
0
0
0
0
0
0.12
0
-
Cadaverine
Tyramine
Putrescine
Amino acids*
0
0
0.033
0
0
0.072
-
0
0
0
0
0
0
0
0.445
0
0
0
0
0
0
0
0
0
0
0
4.02
0
0
0
0
0
0
0
0
0
15.64
0.115
0.121
0.362
A n a ly s is 1
( m g /1 0 0 g )
0.452 0.1
4.31 0.22
0.355 0.03
0.519 0.127
1.953 0.32
0.986 0.1 9
0.219 0.0 9
2.643 0 .1 7
2.912 0.12
1.847 0.22
A n a ly s is 2
( m g /l0 0 g )
0.704 0.11
2.95 0.38
6 .6 8 0.58
3.15 0.54
2.28 0.9 6
1.71 0.33
0.273 0.023
1.892 0.14
0.633 0.03
0.453 0.05
A n a ly s is 3
( m g /1 0 ( ) g )
0.539 0.02
0.281 0.04
8.522 0.71
1.758 0.12
6.314 0.6 7
5.906 1.31
3.071 0.2 9
4.073 0.62
0.375 0.05
0.119 0.04
*The numbers indicated after the cheeses represent which outlet they were
obtained from.
45
200
Level
Cheese
Spike 1
Spike 2
Spike 3
47
223.2
238.3
230.2
100
90.4
96.5
93.2
(ppm)
Recovery
%
46
Figure 3.2 Bar graph of Histamine Levels in Irish Cheeses over three-week period.
H i s t a m i n e L e v e l s in I ri sh C h e e s e s
Analysis 1
(-7 d a y s )
An a lys is 2
( e xp i r y d a t e )
Analysis 3
(+7days)
3 .3 .3 H i s t a m i n e A n a l y s i s i n c a n n e d t u n a
Two brands o f canned tuna were chosen for the survey. Samples were obtained
as with the cheese survey from two different retail outlets to observe batch to
batch variations and analysis was carried out in triplicate.
Canned tuna was tested on opening (Time =0) and then left in storage at room
temperature and at refrigeration temperatures to decompose. Further analysis
was conducted at daily intervals in triplicate.
F i g u r e : 3 .3 A n a l y s i s o f H i s t a m i n e l e v e l s o v e r t i m e o n C a n n e d T u n a a t
R o o m te m p e r a tu r e s to ra g e fro m O u tle t 1.
48
Figure 3.4 Analysis o f Histamine levels over time on Canned Tuna at Room
temperature storage from Outlet 2.
1
0
1 i
12 24 36 48 eo 72 84 96 108 120
OC
Brand2
- -a- - B r a n d i
0
0
12 24 36 48 60 72 84 96 108 120
24
48
96
iiTBanBrcpenmg(roii^
T a b l e 3 .6 S p i k i n g a n d R e c o v e r i e s o f H i s t a m i n e i n T u n a
A 200ppm histamine spike was added to three turn samples prior to homogenisation
and extraction and the percentage o f histamine spike recovered was determined by
fluorimetiy.
Fresh
Spike 1
Spike 2
Spike 3
tuna
Fluorescence
71.82
122.33
139.44
124.2
Level
157
351.0
399.0
357.5
100
97.5
121
99.9
(ppm)
Recovery
50
Histamine (mg/100g)*
Potato (a)
Potato (b)
Carrot
Apple
*(Analysis in triplicate)
and
hours.
Differential media
Proteus vulgaris
Escherichia coli
51
F i g u r e 3 .7
E ffe c tiv e n e s s o f th e io n - c h ro m a to g r a p h y c o lu m n in r e m o v in g h is tid in e
fr o m c u ltu r e flu id s a m p le s .
-pre. column
post column
Histidine(ug/ml)
F i g u r e 3 .8 : G r o w t h C u r v e f o r Providencia retgerri 8 6 5 i n u n s u p p l e m e n t e d a n d
h is tid in e s u p p le m e n te d L B b r o th .
QoAlhQivefcrP. reM
gai inLBandLBtistidre
suJementediTBcfa
3
25
-X'
15
-UsHdB
05
***0*x**
0
-h H4
I - - -I
H- ~h +H -h
8 10 12 1 4 1 6 1 8 2 0 2 2 2 4 2 6
T m e(hoifs)
52
Histamine (mg/100g)
Tuna (a sample)*
88.4
Tima (b sample)*
6 8 .8
+ DAO (5min)
19.1
10.7
+ DAO (overnight)
0 .0
53
3.6 Discussion
Taylor et al. (1978b) showed water saturated n-butanol to be the best organic solvent
for the extraction of histamine with a partition coefficient o f 33. However Shore,
(1971b) suggested that in samples where there are high histidine levels like in the case
of canned tuna, n-butanol-chloroform (3:2) should be used. However these solvents
only have a partition coefficient o f 0.04 and may leave the majority o f histamine in
the aqueous phase and not the desired organic phase. Table 3.1 shows that some
histidine was carried through into the organic phase (up to 4%) and displayed some
residual fluorescence. The putrificative amines cadaverine and putrescine were
analysed in the same way as they are associated with decomposition in fish and may
be a source o f interference. A solution o f these amines along with glutamine and
tryptophan gave an average recovery o f 109%, again indicating that some o f these
substances were not totally excluded from the butanol phase. However these
interferences were greatly decreased by the extraction steps and should not interfere to
any great extent in the food assays.
The fluorometric assay is based on the reaction o f o-phthaldiadehyde with amines at a
high pH. Upon acidification all complexes are dissociated with the exception o f the
acid stable fluorophore OPT-histamine complex. Therefore it should be possible to
observe the total amine levels by eliminating the acidification step and comparing
levels to an acidified sample having only histamine fluorophores. This result would
again demonstrate whether any other amines are being extracted along with histamine
in the procedure. It is apparent from Table 3.2 that there was a difference in levels o f
upto 3 ppm (0.3mg/100g) between the acidified and unacidified samples which is the
amount o f interference coming through the butanol extraction and complexing with
the OPT.
A broad linear range (0-10 p.g/ml) was achieved for histamine with fluorescence
detection (Figure 3.1) and both cheese and tuna samples fell within this range
following dilutions in the sample preparation stages.
Table 3.3 lists a range o f amines and amino acids subjected to fluorescence detection
after butanol extraction in comparison to histamine. Tyramine and the putrificative
amines putrescine and cadaverine showed little or no fluorescence even at very high
concentrations. The amino acid solution showed some activity but only at high
concentrations
(1 0 0
54
0 .1
55
room temperature. From figures 3.3 and 3.4 an increase in histamine levels can be
seen. Both brands purchased from outlet 1 show similar trends with histamine
increasing at a steady rate over time. The brands from outlet 2, show histamine
remaining constant up to 24 hours followed by an increase and further levelling off
after 72 hours. It must be noted that during the course o f the study the brine, which
normally acts as a preserving agent, was removed from the tuna samples thus
allowing spoilage to occur more rapidly. The effect o f storage temperatures is very
noticeable from the comparison o f the trends in figures 3.5 and 3.6 to the previous
figures. Refrigeration temperatures slow down the rate o f decomposition and this can
be seen by the small increases in histamine levels. Temperature is critical for the
growth o f micro-organisms and cold incubation conditions are usually good inhibitors
o f growth. A net increase o f 10 mg/100g occurs at 4C storage compared to an
increase o f 20-30 mg/100g at room temperature storage. These results verify the
critical need for cold storage temperatures o f fish upon harvesting on board fishing
vessels and also in prolonging their safe shelf life.
Recoveries of histamine from the fish flesh were very efficient with an average
recovery o f 106%.
The number o f microbial species with histidine decarboxylase activity within the
microflora population of a food sample is relatively small in size. A screening process
involving the utilisation o f the differential plating medium (Niven et al, 1981), would
be required to distinguish producing from non-producing strains. For the purpose o f
the experiment three bacterial strains were screened for decarboxylase activity, a
known producer, Providencia retgerri 865 obtained from the National Collections of
Industrial and Marine Bacteria Limited (NCIMB) and two unknowns Proteus valgaris
and Escherichia coli D H 5a from DCU stocks.
Positive growth o f colonies on the differential media was represented by purple
colonies growing on a yellow media background indicating that histamine has been
produced from the histidine substrate. This was achieved for the Providencia retgerri
865 strain at a dilution o f 10
strains even at lesser dilutions (Table 3.8), while decreased dilutions for the
Providencia strain resulted in a complete colour change o f yellow to purple for the
whole plate.
56
Once the histamine producing strain had been identified it was decided to investigate
the level o f histamine production over a time period by Providencia retgerri 865. The
bacterial strain was grown on two types o f media, ordinary LB broth and LB broth
supplemented with histidine to act as a substrate for decarboxylase activity. A growth
curve was constructed in order to observe the different phases o f growth especially
the exponential phase where maximum activity takes place. The lag phase proceeded
for about 2.5 hours, which is the time it takes for the micro-organism to adapt to its
surroundings. This was followed by exponential growth until
period o f maximum activity and growth where the rich supply o f nutrients in the
media is utilised. Eventually the nutrients begin to be exhausted and a balance is
achieved between growth and cell death where a plateau in growth is obtained called
the stationary phase.
The bacteria were grown overnight in 5 mis o f LB broth and transferred to two culture
flasks containing LB broth and histidine supplemented LB broth at a dilution o f
1/100. At regular intervals optical density measurements were taken to observe
growth and also culture extracts were subjected to the fluorometric assay according to
the procedure by Behling and Taylor, (1982). This involved the use o f an ion
exchange column to remove histidine and figure 3.7 shows the extreme effectiveness
o f the technique. Figure 3.8 clearly shows the production o f histamine by Providencia
retgerri 865 by decarboxylase activity on the histidine substrate as compared to the
lack o f production when the substrate is absent. The availability o f histidine
supplement in the medium did not influence growth rate.
A rapid increase in histamine levels occurs in the early hours o f growth (0-6 hours)
corresponding to exponential phase on the growth curve where maximum activity is
expected with increased growth. Histamine production levels off once stationary
phase is reached with a level o f ju st over 200 ppm produced after 24 hours. This level
corresponds to 1.8 ^moles/ml. Each o f these growth stages can be seen from figure
3.9. A similar level was produced by Citrobacter freundii T3 after 24 hours at similar
incubation temperatures with levels continuing to rise on further incubation. Tuna fish
infusion broth was used as the medium for growth in this case and based on the
weight o f tuna fish used to prepare the broth, it was calculated that a histamine level
o f 2.5 |amoles/ml was equivalent to 50 mg/100g, which is the FDA Hazard action
limit for histamine in tuna. (Behling and Taylor, 1982).
57
Although a different medium was used in our studies and may not be a true
representation o f conditions in tuna fish flesh, the species Providencia retgerri 865
was observed to produce less histamine than some o f the major producers like
Klebsiella pneumoniae and Proteus morganii. However like these bacteria it was
capable o f producing histamine in a short period o f time which is o f concern as
although some fish may look unspoiled in appearance, they m ay contain significant
levels of histamine.
Table 3.9 shows the ability o f the enzyme diamine oxidase to successfully break down
histamine. In this study histamine was extracted from tuna samples according to the
method described by Taylor et al. (1978b). Diamine oxidase (20mg/ml) was added to
the extracted histamine and incubated at 37C for various time periods (5 min, 15 min
and overnight). A clear reduction in histamine levels resulted over time with the
complete elimination o f histamine after overnight incubation. Increasing the
concentration o f the commercial enzyme or using a more purified form o f the enzyme
would result in decreasing the time to totally eliminate the histamine levels. This has
significant implications in the development o f enzyme based sensors for histamine
detection (Bouvrette et al. 1997). Obviously an enzyme that can breakdown histamine
completely in the shortest time in order to detect one o f the breakdown products is
essential to the detection system.
58
Chapter 4
HPLC Analysis
59
4.1 Introduction
High performance liquid chromatography has been a very popular method for
the analysis o f biogenic amines (Gouygou et al. 1987, Vall and Malle, 1996,
Vall and Gloria, 1997) and the method o f detection has varied from UV to
fluorescent detectors. Biogenic amines with the exception o f the aromatic
amines, do not possess any chromophore or fluorophore, which means that they
have to undergo derivitisation to render them active.
Two methods o f HPLC analysis were employed for the determination o f
histamine. The first method involves benzoylating the amino acid groups on the
amines with benzoyl chloride followed by separation in a methanol:water
gradient coupled with UV detection. (Yen and Hsieh, 1991). However due to
the failure o f this method to successfully detect histamine a second method
according to Wei et al. (1995) was investigated. This method was based on
derivitisation with dansyl chloride, separation in acetonitrile: methanol:water
and UV detection. It is capable o f analysing seven biogenic amines
simultaneously with good sensitivity and selectivity.
Once the method had been successfully adopted and optimised the ability to
identify and quantify histamine in both standards and food samples was
examined. The extent o f the correlation between the fluorimetric method and
the HPLC method for analysis o f histamine in food samples was examined.
60
4.2 Results
The results from derivitisation with benzoyl chloride were deemed to be
inconclusive and shall only be discussed (section 4.4 below).
After derivitisation with dansyl chloride histamine standards were separated on
a C i 8 column and analysed by UV detection. A number o f peaks emerged and
based on the linear increase in peak area corresponding to an increase in
histamine concentration the peak at 11.5 minutes was selected as a possible
candidate peak for histamine. The peak was also observed in decomposed tuna
samples. The sensitivity of the detection method was established by repeated
analysis o f a lower standard range. Eventually a range o f 0.066 to 0.33|ag/200(j.l
was determined (Figure 4.3) which corresponded to a range o f
to 40 mg/lOOg
in food samples after allowing for the dilutions in the extraction procedure
(Taylor et al. 1978b).
Si
4.00E+07
it
|
S 3
<
0)
3.00E+07
2.00E+07
1.00E+07
O.OOE+OO
0
100
250
500
61
750
1000
Figure
4.3
Standard
Curve
for
determination
62
Dansylated
Histamine
by
HPLC
63
65
Figure 4.6 C hrom atogram showing histam ine peak in (a) tu n a and (b)
tuna treated with diam ine oxidase.
66
Sample
Histamine
(mg/1 OOg)
HPLC Analysis
*Tuna 1
Histamine
(mg/1 OOg)
Fluorometric
Assay
90.38
*Tuna 2
91.22
96.04
Spiked Tuna*
263.46
121.04
89.88
4.4 Discussion
The chromatogram from the analysis o f histamine derivitised with benzoyl
chloride showed far too many peaks so it was decided to try to improve the
extraction technique in order to achieve a cleaner chromatogram. Derivitisation
was allowed to proceed for
phase was extracted 2 x 2 times with diethyl ether. Three main peaks were now
observed, 6.08, 7.08 and 7.39 minutes. A blank was run and the only significant
peak occurred at 7.4 minutes indicating that this is not an analyte peak but
probably due to benzoyl chloride.
The most significant peak resulting from histamine standard injections occurred
at 7.01 minutes and a standard curve (Figure 4.1) indicated that this peak was
that o f histamine with a R 2 value o f 0.9396 which is indicative o f linearity. It
was thought that the linear response might have been improved by using an
alternative to diethyl ether in the extraction, as volatile losses were thought to
have been a potential source o f error. Hexane was tried as an alternative but
67
proved to be a very poor extracting solvent with respect to the benzoyl ated
histamine. Butanol was considered but owing to its high boiling point (117C) it
would not have been compatible with the evaporation step. As no substitute was
found, diethyl ether was used for further investigations. It was found that many
o f the interfering peaks could not be successfully removed so it was decided that
it would be more advantageous to achieve good separation in the presence o f
these interferences than to eliminate them from the analysis altogether. The poor
separation o f the histamine peak at 7.08 minutes was slightly improved by
making the mobile phase more polar but two shoulders could still be observed
on the analyte peak (Figure 4.2). Increasing the polarity o f the mobile phase
further did not improve the separation. The pH o f the mobile phase was adjusted
to pH 4 and the gradient run was changed such that the program proceeded
linearly to 70:30, methanol: water. This improved separation but baseline
resolution was not being observed. The mobile phase was increased to 65:30:5,
methanol:water:acetonitrile and the pH dropped further but no improvement in
resolution occurred and peak tailing was becoming evident. It was decided that
baseline resolution would be difficult to achieve. The method could possibly be
used if peak heights were used, but problems may be encountered whereby the
interference peaks may swamp the signal at the retention time o f histamine. This
could result in a very high detection limit. Quantitation would be difficult based
on peak areas owing to the poor separation achieved.
Analysis o f dansylated amines proved to be less complicated and more accurate,
but analysis time was prolonged due to the overnight derivitisation procedure.
Chromatograms were clearer with relatively few peaks observed. Figure 4.4
shows the range o f peaks obtained from the analysis o f histamine standards and
the peak at 11.5 minutes was postulated to be that o f histamine. From the
injection of a range o f histamine concentrations a standard curve was
constructed based on a retention time o f 11.5 minutes and a linear response was
found with a R 2 value o f 0.9967. This was not conclusive evidence that the peak
at 11.5 minutes was that o f histamine, as an increase in peak area was observed
with the peak at 9.5 minutes. Spiking a tuna sample with a known amount o f
histamine would lead to an increase in peak area in the identifying peak. From
Figure 4.5 a notable increase in height/area is seen in the peak eluting at 11.5
68
minutes while the peak at 9.0 minutes remains unchanged. Further confirmation
is achieved through treating a tuna sample with the enzyme diamine oxidase,
which due to its activity on histamine results in a corresponding decrease in
peak height at 11.5 minutes. (Figure 4.6). It is interesting to note that a decrease
in peak height is also associated with the 9.0-minute peak, that aids in its
identity. As it is subject to deamination by diamine oxidase, it therefore must be
a diamine, which has passed into the organic layer during the extraction
procedure. The figures in Tables 3.1 and 3.3 that identify the possibility o f other
amines or amino acids being selected to some degree by the extraction
procedure suggest that the identity o f the diamine may be cadaverine or
putrescine. However the peak is also present in chromatograms o f histamine
standards which exclude the possibility o f other diamines being present unless
contamination o f the stock histamine has occurred which is unlikely. A new
stock o f histamine was analysed with the same resulting chromatograms. The
peak therefore is likely to be due to some component in the extraction or
derivitisation procedure. A blank was derivitised according to procedure and the
resulting chromatogram showed a large peak at 2.5 minutes, this is either the
dansyl chloride itself or 5-dimethylaminonaphthalene-l-sulfonic acid, which is
a hydrolysis product o f dansyl chloride. (Hui and Taylor, 1983).
Table 4.1 compares histamine levels in tuna as detennined by the two methods.
The methods compare favourably with a 0.56% error between methods for
sample 1 and a 5.02% error between methods for sample 2. However, although
the fluorometric assay confirmed the spike with 86.5% o f the 200mg detected
the HPLC method failed to do so to such an extent. This may be as a result o f
the spike being much greater than the linear range which measures from 0-4.2
mg/lOg i.e.: 0-4.2mg in the 100ml methanol extraction step where the spike is
added. This overloading would account for the non-detection o f the spike.
Both methods work well in this study o f histamine but the HPLC method would
prove to be more beneficial elsewhere as it has the ability to simultaneously
detect several biogenic amines. In terms o f analysis time the fluorometric assay
has a distinct advantage over the HPLC method which requires overnight
derivitisation. However the assay does not provide rapid analysis o f analyte or
high sample throughput needed in the food industry.
69
Chapter 5
Histamine Detection by Novel Methods
70
5.1 Introduction
Nitrophenylazophenol cailx[4]arene belongs to a group o f compounds called
chromogenic calixarenes that have the property o f upon complexation with lithium
and to a lesser extent sodium, in the presence o f a base undergo a change in
absorption. A colour change o f yellow to red corresponding to a wavelength shift
from 380 to 520nm. Colour generation arises from the deprotonation o f the acidic
chromophore (-COH) attached near the ligand polar cavity. A proton acceptor is
required for the colour formation, which cannot procced just by the formation o f the
metal ligand complex itself.
71
By complexing the metal with the ligand it is now possible to determine the presence
o f a base by observing the generation o f a colour change (equation
2 ),
which is the
72
Figure 5.3 The absorbance spectra o f the PVC coated fibre with the calixarene
for the addition o f different ammonia concentrations.
73
5.2 Results
A crude sensor based on the amine oxidase-caiixarene detection system was designed.
An ammonia source (the reaction o f diamine oxidase and a substrate or pure
ammonia) was enclosed by a gas permeable teflon seal and the calixarene complex
was immobilised onto a filter disc and placed the other side o f the teflon membrane.
The whole system was enclosed with non-permeable parafiim.
Para film
C a iix a re n e
.Substrate + Enzym i
Teflon.
Slide
pH
Moisture
Heat
Time
Colour
8.4
Yes
\T_
IN O
yellow
8.4
No
No
yellow
10.3
No
No
yellow
10.3
Yes
X T_
XNU
15 mins
Red
10.3
No
Yes
45 mins
Orange
10.3
Yes
Yes
15 mins
Red
74
Control
Colour
Aqueous ammonia
Red
Histamine only
Yellow
Yellow
T a b l e 5 .3 L e v e l o f h i s t a m i n e d e t e c t a b l e b y c a l i x a r e n e .
Histamine
Time
Colour
0.5 mM
12 hours
Orange
0.25 mM
48 hours
Red/orange
0.1 mM
48 hours
Orange
0.05 mM
48 hours
Orange
0.01 mM
Yellow
F i g u r e 5 .5 C o l o u r G e n e r a t i o n o n C a l i x a r e n e d o p e d f i l t e r d is c s .
B . G aseo u s b ase p re se n t
Colour
Red
oxidase
Providencia retgerri 865 - Diamine
Red
oxidase
Histamine was extracted from cheese samples according to Taylor et al. (1978b) and
treated with and without diamine oxidase. The samples were applied to the crude
sensor as shown in figure 5.4. In most samples the red colour was generated
regardless of diamine oxidase outlining the difficulties posed by background
ammonia. However in one cheese sample colour was only seen after the addition o f
the enzyme which indicated that the system detected ammonia as a result o f the
oxidation o f histamine in the food sample.
5.3 Discussion
The reaction o f an amine with an oxidase enzyme yields a product o f ammonia and in
this case histamine is reacted with diamine oxidase. The chromogenic calixarene,
which has been immobilised onto a filter paper disc, undergoes a colour change from
yellow to red on exposure o f ammonia gas. The reaction between enzyme and
substrate produces ammonia but in the aqueous phase and so in order to cross the
teflon barrier the ammonia must be encouraged to enter into the gaseous state.
Ammonia was encouraged to cross the membrane in a number o f ways. Firstly,
knowing that ammonia is a volatile base, heat was applied, then moisture was added
to the calixarene disc on the opposite side o f the reaction to force ammonia across and
also by increasing the pH the NH 4 molecule is deprotonated forming NH 3 gas.
Firstly no colour change was observed at a pH o f 8.4 even at 44C and the presence o f
moisture. The pH was raised to 10.3 for the remainder o f the experiment, as it proved
76
77
activity would yield an ideal microorganism to test with the sensor system. However
the floral content in food is not o f a homogenous nature and may include several
microorganisms that possess this urease activity. This background level o f ammonia
present in the food matrix represents a difficulty in the detection o f ammonia from
amine oxidation.
The solution to the problem is to use a differential measurement o f histamine in food,
firstly without an enzyme added which would represent any background ammonia and
secondly after addition o f an oxidase, so only ammonia produced as a result o f the
enzyme activity on amines present would be detected. The experiment described here
includes the lengthy extraction procedure in isolating histamine so this would not be a
solution to a rapid and efficient method o f analysis. For an ideal detection system the
enzyme would have to be applied directly to a food sample and the evolution o f
ammonia monitored from a difference measurement.The system also would have to
be in an enclosed environment as the calixarene is very sensitive and may be
influenced by contamination, a problem experienced in one o f the laboratories.
It was found that the specific activity o f commercial enzyme used in the study was
very low, 0.14 units/mg solid (1 fimol putrescine oxidised/hour at 37C). Using the
more conventional unit definition (pmol/min), the specific activity would be 0.0023
units/mg solid. Male et al. (1996) cited that for the development o f an amperometric
biosensor for diamines with a satisfactory response, sensitivity and detection limit, the
enzyme must have a specific activity o f about 1 unit/mg solid. This necessitates the
need for purification of the enzyme from a source such as porcine kidney or the
fungus Aspergillus niger. Already in existence are several biosensors capable o f
measuring histamine either by detecting the product o f an amine oxidase reaction,
hydrogen peroxide (Male et al. 1996), measuring the rate o f oxygen consumption by
the same enzyme reaction (Ohashi et al. 1994) and detecting gaseous trimethylamine
directly using the nitrophenylazophenzl calix[4]arene (McCarrick et al. 1994).
The detection system used in this study is based on the same calixarene used by
McCarrick et al. (1994) but is used to detect another product o f the amine oxidase
reaction with biogenic amines, ammonia. The sensitivity o f the calixarene offers great
potential in a detection system and with more investigations may be eventually used
for histamine detection in the food industry.
78
References
79
and
high-pressure
liquid
chromatography.
Journal
of
Bouvrette, P., Male, K.B., Luong, J.H.T. and Gibbs, B.F., 1997,
Amperometric biosensor for diamine using diamine oxidase
purified from
80
S!1;&&2ik blllij
Gao, C.X., Chou, T.Y., Colgan, S.T. and Krull, I.S., Dorschel, C. and
Bidlingmeyer, B., 1988,
New Polymeric Benzotriazole Reagents for Off-Line Derivatizations o f Amines
and Polymines in HPLC, Journal o f Chromatographic Science, 26, 449-457.
81
Fluorometric
Assay
of
Histamine:
Condensation
with
O-
in
octadecyl
silica
columns
with
9-Fluorenylmethyl
66,
853-857
Amines
as
Indicators
of
Seafood
82
Freshness,
Lebensmittel.-
Karube, L, 1992,
Biosensors Handbook o f M easurement Science, vol. 3, Elements o f Change.
193-201. Publishers. Willey, Edited by Sydenham, P.H. and Thon, R.
Amines
in
Foods:
Effects
of
Histamine,
Tyramine
83
and
Male, K.B., Bouvrette, P., Luong, J.H.T. and Gibbs, B.F., 1996,
Amperometric Biosensor for Total Histamine, Putrescine and Cadaverine using
Diamine Oxidase, Journal o f Food Science, 61, 1012-1016.
185-239.
Mondovi, B., Rotilio, G., Costa, M.T. and Agro, A.F., 1971,
Diamine Oxidase (Pig Kidney), Methods in Enzymology, vol. XVII part B, 735741 Academic Press, Edited by Tabour, H. and Tabour, C.
84
86
Sumner, S.S., Speckhard, M.W., Somers, E.B. and Taylor, S.L. 1985,
Isolation o f Histamine-Producing Lactobacillus buchneri from Swiss Cheese
implicated in a food poisoning outbreak, Applied and Environmental
Microbiology, 50,1094-1096.
Taylor, S.L., Gutheretz, L.S., Leatherwood, M., Tillman, F. and Lieber, E.R.,
1978a,
Histamine Production by Food-Borne Bacterial Species, Journal o f Food Safety
1 173-178.
87
88